Configure Breeze Agent Prompts with the Hubspot MCP Client
In this guide, you will learn how to configure Breeze agents with the Hubspot MCP client so your AI workflows can safely call tools, use context from your CRM, and generate accurate responses for sales, service, and marketing teams.
The steps below follow the official configuration options, expanding them into a practical how-to article you can use when setting up your own environment.
Understanding the Hubspot MCP Client Integration
The Hubspot MCP client connects your AI agents to your CRM data and tools, using a Model Context Protocol (MCP) integration. Each agent can call specific tools and receive structured responses that are controlled by prompts you define.
Before configuring any prompts, make sure:
- Your Hubspot account and CRM objects are already set up.
- You have access to the Breeze agents configuration area.
- You know which tools or actions the agent is allowed to call.
Core Concepts for Hubspot MCP Prompt Configuration
When working with the Hubspot MCP client, prompts rely on several core concepts:
- Tools: Actions or data-fetching operations exposed by the MCP client.
- Context: Information about the user, CRM records, or conversation that should influence responses.
- System instructions: High-level rules that define how the agent should behave.
- Prompt templates: Reusable structures that combine instructions, context, and user input.
Each Breeze agent can have different prompt setups, which means you can tailor agents for support, sales, or operations while still using the same Hubspot MCP client integration.
Step-by-Step: Enable Tools for Your Hubspot Breeze Agent
To allow a Breeze agent to use the Hubspot MCP client tools, follow these general steps in your configuration interface:
-
Open the agent configuration panel.
Navigate to the section where tools or integrations are managed for the agent. -
Select the Hubspot MCP client.
Enable the integration so the agent can see the tools exposed by the MCP client. -
Choose which tools the agent can call.
For example, you might enable tools to:- Search CRM contacts.
- Retrieve deals or tickets.
- Update properties or records.
-
Save your changes.
Confirm that the tools are active for this specific agent profile.
This separation ensures that only the relevant Hubspot tools are available, limiting the surface area of what the agent can do and keeping your workflows secure.
Designing System Instructions for Hubspot Agents
System instructions control how the agent should behave and how it should use Hubspot data. Effective instructions should be clear, concise, and robust against ambiguous user queries.
Best practices for Hubspot system prompts
- Describe the agent’s role and objectives in plain language.
- Specify when to call tools and when to answer from existing context.
- Explain how to handle missing or incomplete CRM data.
- Define tone and style requirements (for example, professional and concise).
Example structure for system instructions:
- State that the agent is a CRM assistant working with the Hubspot MCP client.
- Explain that it must only perform actions using approved tools.
- Require that it summarizes tool output in user-friendly language.
- Instruct it to ask clarifying questions when data is not sufficient.
Configuring Context for Hubspot MCP Prompts
Context tells the Breeze agent what it should pay attention to when generating responses. With the Hubspot MCP client, context may include CRM records, conversation history, and metadata from your workflows.
Types of context you can pass
- User context: Name, role, account, or permissions.
- Record context: Associated contact, company, ticket, or deal IDs.
- Interaction context: The last user messages and agent responses.
- Business rules: Product-specific rules you want the agent to respect.
In your configuration interface, you typically specify which pieces of context are available to the agent. Ensure that sensitive data is excluded or masked as needed before being passed through the Hubspot MCP client.
Context safety and reliability
To keep responses reliable and safe:
- Only expose the minimum context required for the task.
- Validate IDs and record references before the agent calls tools.
- Limit free-form notes if they may contain sensitive information.
- Document what context each prompt or workflow depends on.
Building Prompt Templates for Hubspot Workflows
Prompt templates help you standardize how Hubspot-related tasks are executed. Instead of writing a new prompt for each use case, you can design reusable structures.
Common template elements
- Instruction block: Tells the agent what to do.
- Context block: Inserts CRM data or workflow variables.
- User query block: Represents the customer or internal user request.
- Tool guidance block: Explains which tools should be used and in what order.
When configuring templates in the Breeze interface, map each block to the correct variables so the Hubspot MCP client receives structured prompts each time the template runs.
Testing Hubspot MCP Client Prompts
After creating prompts and templates, testing is essential to confirm that the Hubspot MCP client behaves correctly and only uses the tools you intended.
Checklist for testing
- Run sample user queries for each main use case.
- Check whether the agent calls tools appropriately.
- Verify that CRM data from Hubspot is interpreted correctly.
- Ensure that responses follow your system instructions and tone.
- Confirm that no unexpected or unauthorized actions are triggered.
If you see incorrect behavior, refine your system instructions, tighten the context rules, or disable any tools that are not needed for the specific agent.
Maintaining and Updating Hubspot Agent Prompts
As your processes evolve, you may need to update how the Hubspot MCP client is used in your organization. Plan for prompt maintenance as part of your regular operations.
Ongoing optimization steps
- Review analytics or logs for your Breeze agents.
- Identify repeated questions or failure patterns.
- Adjust your templates or context mapping to reduce errors.
- Update system instructions if your business rules change.
- Periodically re-test critical workflows after major updates.
Because prompts directly influence how the agent works with Hubspot data, treat any significant configuration change as a release, with testing and documentation.
Official Hubspot MCP Client Documentation and Further Help
To see the exact configuration fields, technical requirements, and the latest updates, always refer to the official knowledge base article provided by the platform.
You can review the original reference guide here: Configure prompts for Breeze agents with Hubspot MCP client.
For broader strategy support, SEO optimization, and implementation services around AI and CRM workflows, you can also visit Consultevo to explore consulting options.
Next Steps
Once your Breeze agents are configured with the Hubspot MCP client, document your final prompt structures, share them with your team, and monitor results during the first weeks of live use. With clear system instructions, carefully scoped tools, and well-structured context, your agents can reliably support CRM-driven workflows at scale.
Need Help With Hubspot?
If you want expert help building, automating, or scaling your Hubspot , work with ConsultEvo, a team who has a decade of Hubspot experience.
“`
